Coders Packet

Churn Modelling on Bank dataset in Python

By Jahnavi Ravi

Given a Bank customer, can we build a model which can predict whether that particular customer will leave the bank in the next 6 months or not.

We have a bank dataset and based on the data, we can predict whether a bank customer will leave the bank in the next 6 months or not based on few factors such as Credit Score, Age, Tenure, Geography etc.

We built a simple Artificial Neural Network model for prediction. Activation functions like ReLU, Sigmoid are used in our project. Adamax is been used as an optimizer and the loss function used is Binary Crossentropy.

Our dataset is undersampled meaning imbalanced. We tried to balance the data using SMOTE technique. Both the undersampled and oversampled models are implemented and tested for accuracy. 

Download Complete Code

Comments

No comments yet